4.5 KiB
| summary | read_when | ||
|---|---|---|---|
| Use Nillion nilAI privacy-preserving models in Moltbot |
|
Nillion nilAI
Nillion provides privacy-preserving AI inference using blind computation. Your data remains private throughout the entire inference process - nilAI serves models running in a TEE and other techniques to ensure your prompts and responses are never exposed.
Why Nillion in Moltbot
- Privacy-preserving inference using blind computation.
- Reasoning support for complex tasks.
- Tool calling support for function execution.
- OpenAI-compatible
/v1endpoints for easy integration.
Features
| Feature | Support |
|---|---|
| Streaming | Supported |
| Reasoning | Supported |
| Tool/Function calling | Supported |
| Text generation | Supported |
| Vision/Images | Not supported |
Setup
1. Get API Key
- Sign up at nilai.nillion.com
- Navigate to API settings
- Create a new API key
- Copy your API key
2. Configure Moltbot
Option A: Environment Variable
export NILLION_API_KEY="your-api-key"
Option B: Interactive Setup (Recommended)
moltbot onboard --auth-choice nillion-api-key
This will:
- Prompt for your API key (or use existing
NILLION_API_KEY) - Show available Nillion models
- Let you pick your default model
- Configure the provider automatically
Option C: Non-interactive
moltbot onboard --non-interactive \
--auth-choice nillion-api-key \
--nillion-api-key "your-api-key"
3. Verify Setup
# List available models
moltbot models list | grep nillion
# Run an agent turn
moltbot agent --session-id test --message "Hello, are you working?"
Model Selection
After setup, Moltbot shows available Nillion models.
- Default:
nillion/openai/gpt-oss-20b(nilAI Private 20B) for privacy-preserving inference with reasoning.
Change your default model anytime:
moltbot models set "nillion/openai/gpt-oss-20b"
List all available models:
moltbot models list | grep nillion
Configure via moltbot configure
- Run
moltbot configure - Select Model/auth
- Choose Nillion nilAI
Available Models
| Model ID | Display Name | Context (tokens) | Features |
|---|---|---|---|
openai/gpt-oss-20b |
nilAI Private 20B | 128k | Reasoning, Tool calling |
The full model reference is nillion/openai/gpt-oss-20b.
Reasoning Support
nilAI Private 20B supports reasoning, which enables the model to think through complex problems step by step before providing an answer. This is particularly useful for:
- Complex problem solving
- Multi-step reasoning tasks
- Tool/function calling decisions
Tool Calling
nilAI supports function/tool calling, allowing the model to:
- Invoke external tools when needed
- Return structured function call responses
- Handle parallel tool calls
Usage Examples
# Start an agent session
moltbot agent --session-id myproject --message "Help me with my code"
# Use with the TUI (terminal user interface)
moltbot tui
# Send a message via a channel
moltbot message send --target "+1234567890" --text "Hello from nilAI"
# Open the dashboard
moltbot dashboard
Troubleshooting
API key not recognized
echo $NILLION_API_KEY
moltbot models list | grep nillion
Ensure the key is correctly set.
Connection issues
Nillion API is at https://nilai-f910.nillion.network/v1. Ensure your network allows HTTPS connections.
Model not found
The model reference includes the full path: nillion/openai/gpt-oss-20b. Make sure to use the complete reference.
Config file example
{
env: { NILLION_API_KEY: "your-key" },
agents: { defaults: { model: { primary: "nillion/openai/gpt-oss-20b" } } },
models: {
mode: "merge",
providers: {
nillion: {
baseUrl: "https://nilai-f910.nillion.network/v1",
apiKey: "${NILLION_API_KEY}",
api: "openai-responses",
models: [
{
id: "openai/gpt-oss-20b",
name: "nilAI Private 20B",
reasoning: true,
input: ["text"],
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
contextWindow: 128000,
maxTokens: 8192
}
]
}
}
}
}